(2)  Of ranges, transpositions and key signatures 

It seems sensible to deal with the various additional pieces of evidence that have accrued which suggest this ‘symphonic poem’ was indeed intended as a work for brass band.

I claim no special expertise in the brass band field, and I’ve been considerably assisted by conversation and correspondence with John Pickard, and correspondence with Philip Maund, who works at the British Bandsman offices. They have raised several significant points which go a long way to confirming that The Battle Song itself, and probably the transcriptions too, are short scores intended for brass band realisation. Their comments are worth preserving verbatim.

Writing to me on 25 July, without having seen the scores, John Pickard pointed out:

1. You can confirm pretty conclusively whether it was intended for brass band by looking at the range. A composer writing for BB in 1930 would be unlikely to go much above c’’’ (a’’ on Be soprano clarinet) [Well, it so happens that that c’’’ is the highest pitch in The Battle Song short score)]

2. The question of pitch rears its head when dealing with a short score in the BB medium. You may know that the trick employed by many composers (myself included, in the two works I’ve written for the medium) is to write a short score a tone higher than you want it to sound. This means that when you write out in full score you can write all the instruments (except bass trombone and tuned perc.) in C or f, which is much more manageable than messing about with Bb and Eb. This has implications for anyone trying to realise The Battle Song: is it written at pitch or is it already transposed? (Incidentally, has anyone checked the keys of the 15 transcriptions yet? Is it possible that they are transpositions of the works for the benefit of a BB arranger? If the March to the Scaffold is in A minor rather than C minor, that would answer the question as to why Brian made it. Just a thought...)

In fact, the March to the Scaffold remains in Berlioz’ G minor, and in all the cases I have so far been able to determine the transcriptions are written in the same key as their originals. But Philip Maund, in a letter of 16 August, amplified this point:

The keys of the transcriptions might ... yield important clues, but there are many examples of arrangements made directly from the original source, so that the band version sounds a tone lower. In other words, if you found one of the transcriptions was in a different key from the original, that would be a good indication that it was intended for brass band; but the fact that it may be at the same pitch does not necessarily preclude the possibility it was still intended for the band. Another widely recognised rule-of-thumb is that, in general, bands of that time favoured flat rather than sharp keys - obviously there would be exceptions, but it is worth bearing in mind that intonation in sharp keys was more difficult on instruments of the period…

For what it’s worth, the most of The Battle Song is written, like almost all other Brian short scores, without key signature; it opens and closes with music in a key-signature of D flat, and there is a further passage with the (most unusual) seven-flat signature of C flat. Until the possibility arose that this was a brass-band work, these signatures were a rather puzzling sight in a Brian score from as late as 1930-31.

Phil Maund is one of the leading scholars and researchers of the bass band field; yet in the same letter he observes that "…nobody I know had an inkling that Havergal Brian may actually have written a brass band piece". However, he continues:

It sounds as though the provenance of The Battle Song shows fairly conclusively that it was intended for brass band, perhaps even a potential championship test-piece (like Elgar’s Severn Suite (1930) and Bantock’s Prometheus Unbound (1933). At the time R Smith published music exclusively for brass band... Also, Elgar was asked to submit a short score of the Severn Suite before the full score, perhaps as part of a vetting procedure - maybe the same was the case with The Battle Song?

(When in Newsletter 125 1 referred to Henry Geehl assisting Elgar in scoring The Severn Suite I was innocently perpetuating a myth encouraged, it would seem, by Geehl himself, who is thought to have greatly exaggerated his participation. John and Phil have both pointed out that it appears from the most recent research that Elgar did his own band scoring unaided, though Geehl did make a transcription for military band in 1931. They also made me aware that the history and text of the Severn Suite is currently a hotly-debated issue in the brass band world, confirming my suspicion that I would be well advised to stick strictly to Brian!).

After I had provided him with a photocopy of the ms of The Battle Song, Phil Maund wrote to me about the significance of the pitch-range which the short-score reveals. His discussion is worth quoting in extenso (readers unfamiliar with brass band scores should note that all the transposing instruments, even bass tubas, are written in the treble clef):

as far as I can see, the lowest notes are the octave E in bar 13, and an Eb on p5 (2nd stave, b3) The E is the lowest note possible on the Bb bass, written F# below middle C in brass hand (Bb) pitch. The Bb bass at the time did not have a fourth valve (they were not introduced, as far as I know, on the Bb bass until after WW2). The Eb bass, however, could produce an Eb pedal note, and the bottom Eb could also be got with the fourth valve (again, I believe, the fourth valve was introduced on Eb basses somewhat earlier, and I think they were available by 1930.

A ll this suggests to me, at this early stage, that the condensed score (if it is intended for brass band) is at concert pitch and would need to be transposed in terms of key as well as clef. If it is at brass band pitch, these bottom notes would be unplayable. The knock-on effect of this at the top end would mean the top Cs would be fine for the soprano cornet [written A, an octave and a half above middle C] and pushing it a bit for the Bb cornets [written D, two octaves above middle C]. In other words, the pitch range does seem to fit a brass band scenario.

One can perhaps go further: not only does the pitch range fit a brass band scenario, but it is almost inconceivable that Brian would have limited himself to that range if he had intended The Battle Song as a work for full symphony orchestra - which still, of course, doesn’t preclude the possibility of an orchestral realisation subsequent to the writing of the full brass band score.

Some readers may consider this is rather a heavy-weather discussion to establish what I, for one, has assumed from the first: that the short score was at concert pitch, like any other Brian score. But then, to begin with I was thinking that we were faced with an orchestral work. Clearly, when we move into the brass band field, assumptions about pitch cannot be made so lightly. Likewise, I had assumed – and do so still - that Brian would have wanted to preserve the original keys of the various pieces represented in his transcriptions.

I have no direct evidence for this belief, other than a conviction that he was sensitive to key as part of a work’s essential identity. But, as we see, the fact that his short score transcriptions are in the same key as their originals doesn’t rule out the possibility of a brass-band realisation sounding a tone lower. Only because the pitch range of The Battle Song suggests it is at concert pitch, requiring transposition (Phil Maund is careful to stress this is a preliminary impression, though it’s one shared by John Pickard after his own examination of the music), does it seem reasonable to think the transcriptions are likewise at concert pitch.

(3)  The manuscript

The short score is preserved on 12 pages (at first glance they might seem three double foldings) of 28-stave paper, upright format, approximate dimensions 36.3 x 26.5cm. This paper (which carries no manufacturer’s name or watermark) is one that Brian used frequently from the 1930s to the 1950s; all but one of the transcriptions is written on it, and it was used for the short score of the Violin Concerto’s finale. Perhaps significantly (but more likely perhaps not) his preferred medium for the short scores of other major works of the early 1930s was a slightly different, 26 stave paper; Symphony No 2 is entirely on this paper, and No 3 includes just one sheet of 28 stave.

The music occupies ten of the twelve sides, numbered 1-10 from the second side. The twelfth and last side is blank; the first side forms a title-page, boldly written in longhand in a rather greyish ink: ‘Symphonic Poem / The Battle Song / by / Havergal Brian / 1930 / 1931’

Unlike the transcriptions, which appear to be, so to speak, pagines intactae, possibly never opened since they left Brian’s hand, this score shows distinct signs of use. At some time it has been folded in half, horizontally, and there is a resultant crease across the pages. The two middle foldings (music pages 2-9) remain undamaged, but the outer one has split into two separate sheets along the ‘spine’ of the folding. Closer inspection reveals, however, that the original second half of the folding had already been removed: the page containing music p10 and the blank final side is a replacement, part of which remains gummed to the left-hand edge of the title page. This final page is therefore the largest of a number of stuck-over corrected passages which are a feature of the score.

The music is on the whole very neatly written - generally on three staves, though one quasi-fugal passage begins on one and is confined to two for several systems. Brian’s working practice of the time would indicate this manuscript was preceded by a rougher but likewise continuous sketch score in pencil, probably on two staves. (A few leaves of such sketch-scores exist for symphonies 2 and 3, only because Brian used the reverse sides for drafting later works; otherwise pencil sketches of the 1930s are not often preserved).

The ink score exhibits many emendations and corrections. Sotheby’s catalogue described it as written in black ink, but this is (characteristically) incorrect. It was first written - and largely remains - in the dark blue-black common to most of Brian’s scores of this period. Changes and corrections have been made in a mixture of pencil and a different, grey-black ink of lighter tone; this is the ink used for the title-page, the replacement last page, and for all the stick-overs. It seems possible that its tone was once darker and the greyness - not consistent at all appearances - is due to fading; the blue-black is similarly affected.

Pencil notations are plentiful on the first page (where, possibly, working within a key-signature was pausing Brian some notational headaches), and sparser thereafter. A peculiarity is the fact that he has numbered the first bars in pencil. Phil Maund has suggested, and it had independently occurred to me, that this might indicate Brian had progressed thus far in writing a full score - but I cannot recall another instance of him doing this.

L ater on, several systems are marked with a large marginal ‘X’ in the greyish ink - another feature I don’t remember seeing in other Brian short scores. In every case these are systems that have been revised, but whether the ‘X’ signals the need for revision, or the fact that it has been carried out, remains unclear. Systems with stick-overs do not have these inked ‘X’ markings, however. On p9, a sequence of four systems carrying both inked corrections and actual stick-overs have the marginal X’s, but in pencil, not ink. I cannot interpret this except to remark that we may be dealing with two, if not three layers of correction and revision.

Phil Maund has suggested to me that the ‘X’ markings might in fact he in the hand of an editor, who

…thinking some passages were ‘unsuitable’ marked them with an X for Brian to revise, resulting in the emendations and stick-overs. Perhaps they even had the temerity to suggest it needed a ‘bigger’ ending? There is circumstantial evidence which suggest that this kind of heavy-handed interference actually took place.

Though it is almost impossible to assign the ‘X’ markings to a specific hand, I tend to think they are Brian’s own, not least because all the inked ones are in the same ink as the stick-overs, and the stick-overs themselves are consistent with Brian’s revisionary practice of the time. The contemporary short scores of Symphonies 2 and 3, which never had to go before an editor (that of no 3 we know however was certainly shown to Bantock) likewise contain plenty of emendations and stick-overs, some of them very extensive. Still, it remains possible that an editor for R Smith and Co went over the score verbally with Brian, who subsequently added the ‘X’s to help him remember which passages he had been recommended to change.

The score ends at the foot of p10, with no further inscription - none of the details of date and place that Brian vouchsafed in the short scores of Symphonies 2 and 3.

NL127 © 1996 Malcolm MacDonald

BACK TO TITLE PAGE

990424 Havergal Brian - the official website HOME